In this retrospective analysis the predictive value of obtaining a live birth of three IVF indicators has been evaluated in elderly women of couples affected by unexplained infertility: FOI, FORT and OSI do not show a stronger or more informative association with live birth than the components used for their calculation, i.e. the number of oocytes retrieved, the AFC, the number of preovulatory follicles and the FSH total dose. Female age remained the most reliable predictor for live birth in an IVF cycle.
Women of advanced reproductive age remain an open dilemma and a challenge for all clinicians working in the field of assisted reproductive technologies (ART).
Over the years, several attempts have been made, in order to identify surrogate markers of ovarian reserve, which in turn could be markers of IVF outcomes. This would allow to decide for whom treatment is expected to be (cost-)beneficial. However, so far studies on the association between markers such as AMH/AFC and implantation, pregnancy, and/or live birth after assisted conception reported conflicting results [22, 23].
Therefore, several authors have investigated alternative markers that are discussed in this paper, particularly those related to ovarian responsiveness to COS, in order to better predict IVF outcomes. Efficient markers would be of particular interest in subgroups of low prognosis patients defined both by the Poseidon and the Bologna criteria [17, 24].
Alviggi et al. analyzed the predictive role of FOI to assess ovarian sensitivity in hypo-responders patients [10]. The authors concluded that FOI might reflect the dynamics of follicular growth in response to exogenous gonadotropin, better than traditional markers of ovarian reserve. In particular, low FOI values imply that only a fraction of available antral follicles was exploited during COS, suggesting that there might be therapeutic opportunities (increasing the FSH dose and/or adding LH) to improve ovarian responsiveness, and therefore the overall prognosis.
Grynberg et al. discussed the potential use of FORT as a quantitative and qualitative marker of ovarian responsiveness to gonadotropins, and the possible implications for the applicability of the Poseidon criteria [25]. They stated that FORT may reveal impaired sensitivity to FSH and should be used to guide the decision of treatment protocol, gonadotropin, and stimulation doses to be used for hypo-responders.
Recently, the amount of hormone medication needed for each oocyte produced (i.e. the OSI) was investigated in a retrospective cohort study that included more than 1200 women undergoing IVF with FSH/hMG stimulation [26]. Consistently with previous results in younger women with excellent pregnancy potential [7], the OSI was found to be predictive of pregnancy and live birth also in older women with a more unfavorable prognosis [26]. The authors concluded that OSI could be employed in counseling women of advanced age about their reproductive potential, bridging the gap between the purely quantitative aspect of ovarian reserve and the more qualitative approach of ovarian competence.
So far, no study has been published on whether these surrogates offer significantly higher performance than their constituent parameters, particularly in women of advanced age. To be specific: are FORT and FOI more predictive of baseline AFC, the number of pre-ovulatory follicles and the number of oocytes retrieved? As for the OSI, is it more informative than the number of oocytes and the total dose of FSH used?
Despite the limitations of a retrospective study, our results seem to scale back the capabilities of FORT, FOI and OSI to answer the above questions. The combination of multiple indicators of ovarian reserve and ovarian response to COS (FOI and FORT) or in term of FSH administered (OSI) does not seem to be more advantageous than the traditional predictors of IVF outcomes. In addition, most of these predictive factors only become available after at least one IVF cycle was conducted, limiting their usage in clinical decision-making on starting IVF. Thus, there is still a knowledge gap regarding the possibility to predict the oocyte quality, especially using only information that is available before starting COS.
In a cohort of couples with unexplained infertility and advanced female age, only age was found to be a clear predictor of live birth after IVF, further confirming the intrinsic awareness of every IVF expert: the age of the egg is what really matters. This conclusion is in agreement with what has already been reported in previous studies [22, 27, 28].
Indeed, numerous factors are implicated in the final outcome of an IVF treatment: oocyte and embryo quality, endometrial receptivity, woman general health conditions, etc. Probably, only the female age is able to coherently capture all these factors. The impact of age per se seems truly relevant, considering that the correlation with IVF outcome is stronger than that of all markers considered or even combined, in spite of a rather narrow distribution of ages.
Note that FORT, FOI, and OSI can be considered markers that indicate different aspects of the response to COS; we merely showed that their association with live birth after IVF are negligible when considering female age and their constituent parameters.
This result could have more than one explanation. First of all, one should consider that each of the parameters used for the calculation of the three metrics are operator- dependent. The follicle count, which registered a reduction in inter-individual variability with the introduction of 3D technology [29], is still widely evaluated in 2D in most of the centers, as in the present study. Such a variable is inevitably affected by the skill and the accuracy of the operator. Similarly, there is no unanimous agreement on which should be the starting dose of gonadotropins in an IVF treatment [30, 31]. Whereas several algorithms have been developed over the years [32, 33], the starting dose is still widely established on the basis of the operator's clinical sensitivity with respect to multiple parameters (AMH, AFC, age, body mass, previous COS etc.). Obviously, the total dose of gonadotropins administered is partly affected by the initial decision.
Also the number of oocytes retrieved can o be affected by the experience of the operator [34].
In summary, all of these parameters suffer from measurement error, opening up the possibility of adding more noise i.e. (non-)random variability. By adding in a model more variables that are operator-dependent it is not surprising that the accuracy of these surrogates is reduced with respect to the single parameter that underlies it. This could also explain why a simple but perfect measure such as female age seems to be more informative.
The main strength of our retrospective analysis is that it represents the first study to investigate whether a non-linear association exists between the metrics FORT, FOI and OSI and LBR in the same homogeneous population, i.e. women of advanced age with unexplained infertility, thus avoiding all the hypothetical confounding factors deriving from all the other plausible causes of infertility (male infertility, endometriosis, tubal obstruction, etc.). In addition, we carefully assessed the fit of models with the constituent parameters using contemporary statistical methods, adjusting for multiple comparisons and adjusting standard errors for couples receiving multiple cycles.
In conclusion, none of the three predictors for fertility that were proposed in literature (FORT, FOI and OSI) were more predictive than the two ‘components’ characteristics that were used to calculate them. This was in a population of women of advanced age (39 years or above) with unexplained infertility. Nonetheless, in light of a lack of evidence that backs up using the three metrics, we suggest that clinicians and researchers still to using the components themselves (i.e. the number of oocytes collected, the number of preovulatory follicles, AFC and FSH dose) in counselling and prediction modelling.